More parsing
Daniel Riaño
danielrr at eresmas.net
Sat Apr 20 19:18:37 EDT 2002
Clayton S. B. wrote:
>
>I guess that from my eclectic functional point of view, the construction:
>article -> prep (additive, substantive, pronoun) does not really present one
>with any new problems in terms of parsing trees.
Lucky you, Clayton! Maybe it is a simpler problem than I thought, but
I still see a theoretical problem here. It is not a problem of
semantics (the semantics are clear), but an intriguing question for
any coherent, robust theory of syntactic description. There are many
traditional ways to parse the text, of course, and most of them just
beg the question. But if you want to parse the immediate constituents
of the syntagm, you face a few problems with this kind of
construction.
Here are a few alternative trees for the first syntagm (we don't
really need any more text to parse it), and all of them present some
problems. And there are more possible analysis. Have a look at it. I
think you'll visualise it a lot better if you change your display
font to a monospaced (= fixed-width) font, like Monaco or Courier
(Eudora users, click the "I<-->|" button at the top menubar of this
message).
Please note 1: In the following trees, I just give the immediate
constituents' lexical content, not the syntactic tag or the
syntagmatic classification, to simplify things
Please note 2: The main constituent (the nucleus of the syntagms) is
always on the left of the branch.
Please note 3: I know Chomskians would put the preposition at the
left side of the branch: for the sake of the argument we could pass
upon this question, because it is not relevant for this particular
point.
Final note: Sometimes I'll use just the few letters of every word for
the sake of brevity. T1 is the first T=, ktl.
(A) TO\ E)PI/BLHMA TO\ A)PO\ TOU= KAINOU=
/\
/ \
/ \
/ \
/ \
/ \
/ \
/ \
/ \
TO EPIBLHMA TO APO TOU KAINOU
/\ /\
/ \ / \
/ \ / \
/ \ / \
E T1 TO-2 A. TOU K.
/\
/ \
/ \
/ \
TOU K. APO
\
/ \
/ \
/ \
KAINOU TOU
(B) TO\ E)PI/BLHMA TO\ A)PO\ TOU= KAINOU=
/\
/ \
/ \
/ \
/ \
/ \
/ \
/ \
/ \
TO EPIBLHMA TO APO TOU KAINOU
/\ /\
/ \ / \
/ \ / \
/ \ / \
EPIBLHMA TO A. TOU K. TO -2
/\
/ \
/ \
/ \
TOU K. APO
\
/ \
/ \
/ \
KAINOU TOU
(C) TO\ E)PI/BLHMA TO\ A)PO\ TOU= KAINOU=
/\
/ \
/ \
/ \
/ \
/ \
/ \
/ \
/ \
TO EPIBLHMA APO TOU KAINOU
/\ /\
/ \ / \
/ \ / \
/ \ / \
EPIBLHMA TO-1(TO-2) TOU K. APO
/\
/ \
/ \
/ \
KAINOU TOU
(D) TO\ E)PI/BLHMA TO\ A)PO\ TOU= KAINOU=
/\
/ \
/ \
/ \
/ \
/ \
/ \
/ \
/ \
EPIBLHMA APO TOU KAINOU TO-1 (TO-2)
/\
/ \
/ \
/ \
/ \
/ \
EPIBLHMA APO TOU KAINOU
/\
/ \
/ \
/ \
TOU KAINOU APO
/\
/ \
/ \
TOU KAINOU
And many more... So what is the actual parsing used in the parsed
editions of the NT, or the BGreekers proposals? And you, Clayton,
what do you think now?
> > Dear BGreekers,
>>
>> I don't have any parsed text of the NT available, but I am curious to
>> know how the authors parse the SN or SP that is placed formally in
>> dependence of an article in attributive position (I would also like
>> to know how do they parse the article and the resulting tree!), i.e.
>> the SN or SP between the asterisks in:
>>
> > TO\ E)PI/BLHMA TO\ *A)PO\ TOU= KAINOU=* Eu.Luc.5.36
>> H( PI/STIS H( *DI' AU)TOU=* Act.Ap.3.16
>> TH\N A)NA/STASIN TH\N *E)K NEKRW=N* Act.Ap.4.1-2
>>
>> What I am interested in is not in the *tagging* of either the article
>> or the dependent syntagmata, but the *parsing* of its constituents
>> within the whole group. I imagine there is Chomskyan procedure to
>> parse such sentences, and I'd like to know it too, but I am more
>> interested in non-generative approaches. Does anybody know examples
>> of not Indo-European languages with the same construction?
>
>Daniel,
>
>Nice question. Not the sort of thing one can just look up in BDF, Zerwick or
>any of the lesser grammars.
>
>...
>>From a functional perspective one could say that the constituent TO APO TOU
>KAINOU in LUKE 5:36, limits TO EPIBLHMA. So I would parse it the same way
>you would an arthrous adjective or noun in the same slot. How to parse the
>elements of the prepositional phrase itself? Why bother? It performs as a
>unit adjectivally, the parsing of the leaves of the tree does not seem to
>fulfil any useful purpose once the function of the unit is identified. Some
>hard core generative types might take issue with this statement. The
>constituent THN EK NEKRWN in ACTS 4:1-2, limits THN ANASTASIN. Same
>situation, same comments.
>
>The constituent hH DI' AUTOU in ACTS 3:16 limits hH PISTIS. Formally this
>looks like the other two examples, but there are some intriguing semantic
>ambiguities in here that are lacking in the other two examples. Determining
>just exactly what DI' AUTOU means in this context isn't trivial. However,
>not sure this will impact the parsing problem any since I would assume that
>you are parsing syntax functions and not semantic ones. However, as anyone
>who has been through the generative school of thought knows, keeping the
>syntax functions and semantic ones separate is easy to say but difficult to
>implement.
--
~~~~~~~~~~~~~~~~~~~
Daniel Riaño Rufilanchas
Madrid, España
More information about the B-Greek
mailing list